Immediate and delayed (7, 14, and 28 days) free recalls of 40-item lists were studied. The lists consisted either of high-or low-frequency associates of the category names and were presented in either blocked or random order. The delay groups did not differ at immediate recall, but item recall decreased with time up to 14 days. Categorial intrusions increased with delay. List organization as measured by repetitions declined with delay but not as measured by ratio of repetition. The findings are interpreted in terms of loss of knowledge of the list membership of categorial items with time_ Aside from two reports by Brand (1956; Brand & Woods, 1958), there has been little work on the free recall of lists of words over substantial retention intervals. In Brand's studies, recalls were obtained at intervals of 1,2, and 3 weeks following presentation of a list of items from the categories of the Allport-Vernon-Lindzey Study of Values. In these studies, there was an increase in the number of categorial intrusions over the retention interval. However, the value of these results is lessened by certain limitations of method; further , work on free recall has not usually employed value related materials. Consequently, it has seemed worthwhile to repeat Brand's experiments with more standard materials and under better controlled conditions.In free recall of word lists, two findings lend further interest to the study of long-term retention. The first is that, for list lengths in excess of memory span, mean recall is always less than the number of items presented_ The second is that the number of intrusions is very small. Since the words used in the lists are usually familiar ones and are clear instances of the categories employed, a plausible interpretation of these findings is that at recall subjects, while able to produce enough words to match list length, are reluctant to do so when they are uncertain as to the list membership of the words which occur to them (Cofer, 1967, Experiment 2). If this interpretation is valid and if, with increases in retention interval knowledge of list membership becomes less certain, then the number of intrusions should increase. The present experiment was designed to test this derivation which, of course, was confirmed in Brand's data as indicated above. Two additional factors were manipulated, and both could be related to number of intrusions. One is manner of presentation, with the list members presented either in random order or blocked by category. The other is the frequency with which the list items are given in response to the category names in an association test. The retention intervals studied were 7, 14, and 28 days following presentation and immediate recall of the list.
METHODThe experiment followed a 2 by 2 by 3 design, with the conditions being kind of list, presentation orders, and retention intervals.
ListsThere were two lists of 40 words each, comprised in one case of the 10 highest frequency associates to the category names, occupations, weapons, four-legged animals, and ...